STSylvie Treuillet

Sylvie Treuillet

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Sylvie Treuillet has a total of 3 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • 3D Modeling from Uncalibrated Color Images for a Complete Wound Assessment Tool

    by B. Albouy, Y. Lucas, S. Treuillet

    This paper is concerned with the 3D modeling of skin wound using uncalibrated vision techniques for the volumetric assessment of the healing process. We have developed an original approach for matching two color images captured with a free-handled digital camera and generate a semi-dense 3D model. We evaluate the precision of the inferred 3D model by registration to a ground truth on artificial wounds. The method is then applied to volumetric measurements. The clinician requirements of a global 5% precision are overshot as 3% is obtained locally. The best configuration for taking photos lies between 1.2 and 1.5 for distance ratios and between 15deg and 30deg for vergence of the stereo pair. This work is part of the ESCALE project dedicated to the design of a complete 3D and color wound assessment tool using a simple free handled digital camera: a smart solution for massive diffusion in care centers as such very low cost system should be operated directly by nurses.


  • Supervised Tissue Classification from Color Images for a Complete Wound Assessment Tool

    by Hazem Wannous, Sylvie Treuillet, Yves Lucas

    This work is part of the ESCALE project dedicated to the design of a complete 3D and color wound assessment tool using a simple free handled digital camera. The first part was concerned with the computation of a 3D model for wound measurements using uncalibrated vision techniques. This paper presents the second part which deals with color classification of wound tissues, a prior step before to combine shape and color analysis in a single tool for real tissue surface measurements. As direct pixel classification proved to be inefficient for tissue wound labeling, we have adopted an original approach based on unsupervised segmentation prior to classification, to improve the robustness of the labeling step by considering spatial continuity and homogeneity. A ground truth is first provided by merging the images collected and labeled by clinicians. Then, color and texture tissue descriptors are extracted on labeled regions of this learning database to design a SVM region classifier, achieving 88% success overlap score. Finally, we apply unsupervised color region segmentation on test images and classify the regions. Compared to the ground truth, segmentation driven classification and clinician labeling achieve similar performance, around 75 % for granulation and 60 % for slough.


  • Finding Two Optimal Positions of a Hand-Held Camera for the Best Reconstruction

    by S. Treuillet, B. Albouy, Y. Lucas

    This paper proposes an experimental study to find the two optimal positions of a hand-held digital camera for the capture of the geometry and texture of an object. Using our improved 3D reconstruction pipeline based on a semi-dense matching between a pair of uncalibrated images, a layout of twenty-five camera positions is tested in real conditions of image acquisition and also completely simulated. The reconstruction quality is measured by assessing the accuracy of the final 3D structure in accordance with a ground truth. The results provide the optimal capturing layout. Another interesting conclusion is that the accuracy of the reconstruction does not change much in the nearby area around the best position, which enables to the hand-held capture to not strictly respect this configuration.